VScrybe
Download

Voice notes to text,
right on your computer.

Drop in WhatsApp, Telegram or phone recordings and get clean, searchable text in about the time it takes to play them. Nothing is ever uploaded.

Download VScrybe Free for Windows, macOS and Linux · Other downloads
VScrybe is a free app for Windows, Mac and Linux computers.
A VScrybe session of three WhatsApp voice notes, with timestamps beside every transcribed line A VScrybe session of three WhatsApp voice notes, with timestamps beside every transcribed line

Drop in anything.

Pick files, drag them onto the window, or convert a whole folder at once. Every batch becomes a session you can rename, search and add to later.

  • .opus
  • .ogg
  • .m4a
  • .mp3
  • .wav
  • .amr
  • .aac
  • .webm
Three voice notes waiting to be sent for transcription Three voice notes waiting to be sent for transcription

Click a line. Hear it.

Every line gets a timestamp. Click one to jump to that exact moment, then copy the text with or without times, or save it as a file.

A transcript line highlighted while its audio plays A transcript line highlighted while its audio plays

Or just talk.

Press the mic to record a new note. Pause, resume and preview it, then send it to be transcribed like any other file.

Recording a new voice note in VScrybe Recording a new voice note in VScrybe

Speaks your language.

English by default, or pick one of 15 others, or let VScrybe detect it. Accented English is handled well, even with the Fast model.

  • Yoruba
  • Hausa
  • Swahili
  • Arabic
  • French
  • Spanish
  • Portuguese
  • Hindi
  • Chinese
  • +6 more
The language picker open, with English selected The language picker open, with English selected

Fast or Accurate. Your call.

Fast suits most voice notes. Accurate is better with heavy accents and background noise. You can also add your own models from Hugging Face.

Settings showing the built-in Fast and Accurate models Settings showing the built-in Fast and Accurate models

Nothing leaves
your computer.

Other apps send your recordings to a server. VScrybe runs OpenAI's open Whisper models on your own machine, so your conversations stay yours.

Free for every computer.

A small download. The first time you transcribe, VScrybe fetches its speech model once.

Windows

Windows 10 and 11

Download installer

If Windows says it protected your PC, choose More info, then Run anyway. The app isn't code-signed yet.

macOS

macOS 11 or later

Apple silicon (M1 and later) Intel Mac

Not sure which? Apple menu → About This Mac → Chip. First launch: System Settings → Privacy & Security → Open Anyway.

Linux

64-bit · .deb or AppImage

Download .deb AppImage · any distro

The .deb suits Ubuntu, Debian, Mint and Pop!_OS. The AppImage runs on any distro: make it executable and open it.

Questions, answered.

Something else? Use About → Report an issue in the app, or open an issue on GitHub.

Is it really free?

Yes. VScrybe is free and open source under the MIT licence. No trial, no account, no ads.

Does it upload my audio?

Never. Transcription runs entirely on your computer, and your recordings and transcripts are stored there too.

Why does it download 480 MB the first time?

That's the speech model that does the transcribing. It's downloaded once and then works offline. The optional Accurate model is about 1.6 GB.

Which languages does it understand?

English by default, plus French, Spanish, Portuguese, German, Italian, Dutch, Arabic, Swahili, Yoruba, Hausa, Hindi, Russian, Turkish, Chinese and Japanese, or Auto-detect.

Which files can I add?

Voice notes and recordings in .opus, .ogg, .m4a, .mp3, .wav, .amr, .aac and .webm, including WhatsApp and Telegram voice notes.

Your voice notes, finally searchable.

Download VScrybe